1

Journal: Hydrological Sciences Journal

Reference: HSJ-2015-0337

Paper title: Prediction of climate change effects on the runoff regime of a forested catchment in

northern Iran

Authors: Farhad Hajian1 †

, Alan P. Dykes2 *, Bagher Zahabiyoun

3 and Maia Ibsen

4

Author affiliations: 1School of Civil Engineering and Construction (since renamed School of Natural

and Built Environments), Kingston University, Penrhyn Road, Kingston upon

Thames, KT1 2EE, UK.

f.hajian@yahoo.com 2

Centre for Engineering, Environment and Society Research, School of Natural and

Built Environments, Kingston University, Penrhyn Road, Kingston upon Thames,

KT1 2EE, UK.

a.p.dykes@kingston.ac.uk

3Water Resources Engineering Group, School of Civil Engineering, Iran University

of Science and Technology, Narmak, Tehran, PO Box 16765-163, Iran.

Bagher@iust.ac.ir 4

Centre for Engineering, Environment and Society Research, School of Natural and

Built Environments, Kingston University, Penrhyn Road, Kingston upon Thames,

KT1 2EE, UK.

m.ibsen@kingston.ac.uk

†Present address: Department of Civil Engineering, Neyshabur Branch, Islamic

Azad University, Neyshabur, Iran.

Correspondence: *Corresponding author (A P Dykes). Address and Email shown above.

Tel.: +44 (0)208 417 7018

Abstract: The southern coast of the Caspian Sea in northern Iran is bordered by a mountain

range with forested catchments which are susceptible to droughts and floods. This

paper examines possible changes to runoff patterns from one of these catchments in

response to climate change scenarios. The HEC-HMS rainfall-runoff model was

used with downscaled future rainfall and temperature data from 13 Global

Circulation Models, and meteorological and hydrometric data from the Casilian (or

‘Kassilian’) Catchment. Annual and seasonal predictions of runoff change for three

future emissions scenarios were obtained, which suggest significantly higher spring

rainfall with increased risk of flooding and significantly lower summer rainfall

leading to a higher probability of droughts. “Flash floods” arising from extreme

rainfall may become more frequent, occurring at any time of year. These findings

indicate a need for strategic planning of water resource management and mitigation

measures for increasing flood hazards.

Key words: northern Iran, climate change, runoff modelling, weather generator, rainfall patterns

2

Prediction of climate change effects on the runoff regime of a forested catchment in northern

Iran

Abstract

The southern coast of the Caspian Sea in northern Iran is bordered by a mountain range with

forested catchments which are susceptible to droughts and floods. This paper examines possible

changes to runoff patterns from one of these catchments in response to climate change scenarios.

The HEC-HMS rainfall-runoff model was used with downscaled future rainfall and temperature

data from 13 Global Circulation Models, and meteorological and hydrometric data from the

Casilian (or ‘Kassilian’) Catchment. Annual and seasonal predictions of runoff change for three

future emissions scenarios were obtained, which suggest significantly higher spring rainfall with

increased risk of flooding and significantly lower summer rainfall leading to a higher probability of

droughts. “Flash floods” arising from extreme rainfall may become more frequent, occurring at any

time of year. These findings indicate a need for strategic planning of water resource management

and mitigation measures for increasing flood hazards.

Key words: northern Iran, climate change, runoff modelling, weather generator, rainfall patterns

1 INTRODUCTION

There is increasing consensus from different GCMs that as the atmosphere continues to warm,

rainfall intensity is likely to increase over many parts of the world with longer periods of very low

rainfall in between and that increases in rainfall extremes are expected to be greater than the

changes in mean precipitation (Easterling et al. 2000, Wilby and Wigley 2002, SWCS 2003, Kharin

and Zwiers 2005, IPCC 2007). The mean air temperature near the surface of the earth increased by

up to 0.35°C from the 1910s to the 1940s and by as much as 0.55°C from the 1970s to 2007 (IPCC

2007). For every 1°C increase in temperature, the water-holding capacity of the atmosphere will

increase by about 7%, as predicted by the Clausius-Clapeyron equation which defines the vapour

pressure curve for two-phase media (e.g. Venturini et al. 2008). Thus, a warmer climate may be

expected to result in more rainfall but global climate models (GCMs) indicate more complex

effects.

The type (e.g. convective, orographic, etc.), frequency, amount and intensity of rainfall are

all predicted to change. If the frequency of dry days increases due to warming, this does not

necessarily mean that the frequency of extreme rainfall events will decrease, depending on the

threshold used to define such events (Barnett et al. 2006, IPCC 2007, Kundzewicz et al. 2013).

With the increased water vapour in the atmosphere, more frequent and higher intensity rainfall

events are likely to occur in many regions (Douville et al. 2002). High intensity rainfall, particularly

if the total amount of storm rainfall is also very high as is often the case, commonly gives rise to

short-duration, potentially high impact hazards such as flash floods. As a direct consequence of the

changing rainfall patterns, landslides are also expected to increase in frequency. These constitute

additional but related hazards that may be triggered by higher storm rainfall or, in the case of debris

flows, caused by storm runoff in steep mountain streams entraining sediment. There are also signs

that a higher incidence of prolonged droughts may be expected, especially in warmer climates. A

series of Atmosphere-Ocean General Circulation Model (AOGCMs) simulations––run in advance

of a 2008 IPCC report to represent the effects of a warmer climate––indicated a decrease in summer

rainfall in most parts of the mid-latitudes (including Iran) and, thus, a greater risk of drought in

these regions (Bates et al. 2008).

Northern Iran, comprising provinces that border the southern coastline of the Caspian Sea

(Fig. 1), is susceptible to droughts and floods. Planning for improved water management for

drought periods and infrastructure protection for floods is particularly important because it is one of

3

Iran’s major tourist areas. Katiraei et al. (2007) found a trend of increasing annual rainfall totals

over the period of 1960–2001 in the east of Mazandaran Province (significant at p = 0.10) and in

2001 there was a devastating flood in Golestan Province that resulted from a large rainfall event

(>150 mm in 12 hours) of which the first hour was of very high intensity (50 mm h–1

) (Sharifi et al.

2012). Rainfall in the northern part of Iran mainly originates from southerly extensions of Siberian

anticyclones. As high pressure develops over Siberia, elongated ridges of high pressure move to the

north of Iran and pass over the Caspian Sea, receiving heat and humidity from the surface of the

sea. The moist airflow moves south and rises as it hits the northern slope of the mountains, cooling

and condensing to producing orographic precipitation. According to Rasoli et al. (2012), the

Siberian high pressure system has decreased in strength in winter by up to 0.005 mb y–1

over a 61-

year period (1948–2009). Therefore, the airflow from the north to the south of the Caspian Sea

became weaker and transported less moisture, resulting in lower winter rainfall in northern

Iran. However, the pressure of the central area of the Siberian high pressure in spring has increased

by up to 0.059 mb y–1

, producing higher spring rainfall in northern Iran, generally of higher

intensity, thus leading to an expectation of more floods in the southern Caspian Sea coastal area

during the spring season (Rasoli et al. 2012). In parallel with the increasing rainfall, Tabari et al.

(2011) found a significant annual trend of decreasing water levels in 190 observation wells in the

east of Mazandaran Province during 1985–2007, although they did not explain the reasons for this

trend. Possibilities include a higher proportion of convective rainfall which generates a higher

runoff ratio and higher demand for abstracted groundwater from rising immigrant and tourist

numbers.

Figure 1. Location of the Casilian Catchment in Mazandaran Province, northern Iran, showing

locations of hydrometric monitoring stations within the catchment. Valikbon station defines the

outflow from the study catchment.

4

Long-term changes to runoff regimes and groundwater resources have also been observed in

many other parts of the world. For example, Banasik and Hejduk (2012) found a trend of decreasing

annual runoff from a small catchment in Poland over a 48-year period with no apparent cause, in

that there had been no precipitation trend and no human interventions. However, other recent

studies have distinguished the influence of climatic variations from the more tangible evidence of

anthropogenic modification of catchments and increasing utilisation of available water resources.

Chang et al. (2015) found that in the large catchment of the Weihe River in northwest China, a

declining trend in annual runoff was associated with reducing precipitation during 1956-70 but that

after 1970 human activities had twice as much influence on the continuing runoff reductions;

similarly, the contributions of human activities was three times higher since 1990 in the Wei River

basin in China (Zhan et al. 2014). Regionally, it becomes more difficult to explain runoff trends as

there may be complex interactions between different factors in different parts of a region or even

different subcatchments of a large river basin, especially if some rivers show long-term increases in

annual runoff (e.g. in some European countries: Stahl et al. 2010) while in others the runoff is seen

to be reducing (e.g. Li et al. 2014, Jiang et al. 2015), and seasonal changes may be more significant

in magnitude and potential impact than annual trends (Stahl et al. 2010). Even at much smaller

spatial scales, if additional natural factors such as snowmelt or significant groundwater storage

contribute to the overall hydrological response of a catchment, it becomes increasingly difficult to

separate climatic effects from human activities (e.g. Langhammer et al. 2015). Predictions of future

runoff trends resulting from different land use or catchment management scenarios ultimately

require adequate understanding of possible climatic patterns and corresponding runoff responses.

This study addresses the topic from the latter perspective.

The aim of this paper is to use outputs from General Circulation Models (GCMs––often

referred to as “global climate models”) to identify and characterise possible future changes in the

runoff regimes of forested catchments in northern Iran that drain into the Caspian Sea along its

southern coastline. Runoff for future climate scenarios was estimated using a publicly available

catchment runoff model. The working hypothesis is that runoff regimes will show significant

changes compared with the present, with the direction of change (i.e. increasing or decreasing

runoff) varying both annually and seasonally in line with the rainfall trends identified in recent

decades. Implications of the results for future water management are identified and discussed.

2 STUDY SITE

This research focused on the Casilian Catchment (also known in Iran as Cassilian, Kasilian or

Kassilian Catchment) in Mazandaran Province, northern Iran (53°18’ to 53°30’E and 35°58’ to

36°07’N) (Fig. 1). This catchment was selected because the data availability was far greater than for

other areas. Tamab, Iran’s Water Resources Research Organization based in Tehran, collect basic

climate (rainfall and temperature) and runoff (discharge and sediment) data for many catchments in

Iran (Parvardeh 1985). The data used for this research were originally provided by the Head of

Tamab’s Surface Water Section, J. Parvardeh, to BZ for research on Casilian Catchment. The study

area within this catchment is bounded to the north by the Setic and Chartab mountains with heights

of up to 1790 m and to the south by the Mirozad (2700 m) and Golrad (3349 m) mountains. To the

west is the Getoja Mountain (2043 m) and to the east is the Chartab Mountain (1613 m). The main

river drains northwards, eventually discharging into the Caspian Sea. The study area is defined as

the upper 65.7 km2 of the catchment upstream of Valikbon hydrometric station (see below) at 1120

m elevation. Most of the upper tributaries extend to the southern rim of the catchment at around

3000–3300 m elevation. The longest flow path, representing the catchment length, is 17.8 km.

Indicative mean channel gradients are 0.5 for the first two kilometres of the upper tributaries down

the steep escarpment that defines the southern edge of the catchment, then 0.057 for the next 15.8

km. Approximately 80% of the catchment is forested, the upstream (southern) half of the catchment

comprising high quality forest that could theoretically be susceptible to significant deforestation in

5

the future. Most of the downstream area of the catchment has been cleared for agriculture (crops

and pasture) and there are several small villages. The geology of the catchment is dominated by

shale, sandstone, marl and siltstone.

There are two meteorological stations (Darzikola and Sangdeh), one rainfall station (Kale)

and one hydrometric station (Valikbon) in the Casilian Catchment (Fig.1). Sangdeh station, located

at the southeast end of Sangdeh village, records daily rainfall and daily temperatures. Darzikola

station only records daily rainfall data, as does Kale station. The Valikbon hydrometric station, on

the Casilian River, is located near to Valikbon village and records the discharge from the upper

catchment that comprises the study area.

The mean annual rainfall on the Casilian Catchment for the 20-year period 1977–1996 was

estimated to be around 756 mm, using Thiessen polygons (McCuen 1998) formed around the three

rainfall stations. Monthly mean rainfall varies throughout the year from around 48 mm in December

(minimum) to 89 mm in August (maximum). The mean annual runoff from 13 full years between

1980 and 1996 for which complete discharge data exist was 229 mm. The mean runoff ratio for

these 13 years was 0.27. However, the runoff ratio for March and April (61 days) was highly

variable across the 13 years of data, exceeding 1.0 in 1980 (1.34) and 1982 (2.00). We interpret this

as demonstrating a major snowmelt contribution to the runoff regime (after Zampieri et al. 2015).

According to Tamab, snowfall on this catchment is negligible. However, it is likely that the Tamab

data and residents’ comments are referring to snow at Sangdeh and further downstream. As is the

case on the mountains along the northern side of Tehran we think that there is significant snowfall

most winters on the high mountains upstream of Sangdeh where people rarely go and where there

are no monitoring facilities.

3 METHODOLOGY

We obtained the most complete set of catchment hydrometric data for northern Iran that was

available from Tamab, which was for the Casilian Catchment (Parvardeh 1985). The outline

strategy for this research was to use downscaled GCM climate data for three future periods of the

21st century and three global emissions scenarios as the input data for estimating total runoff

volumes from the catchment using a suitable rainfall-runoff model. GCM data were downscaled

using a Weather Generator (WG). The simulations were performed for mean annual conditions and

for seasonal conditions.

3.1 Field data

Daily precipitation and temperature data for the Casilian Catchment were obtained at Sangdeh for

the period 23 September 1976 to 22 September 1997, but with some missing data. Discharge data

from Valikbon were available for the periods 23 September 1979 to 22 September 1987, 23

September 1988 to 22 September 1993 and 23 September 1994 to 22 September 1998. The

incomplete discharge record limited the durations of time periods that could be utilised for this

study. The longest continuous complete record included calendar years 1980–86 inclusive. This 7-

year period was thus designated the “baseline” period and its discharge regime is summarised in

Table 3.

The accuracy and reliability of the recorded field data are not known and cannot be verified.

Three of the authors of this paper (FH, APD, BZ) visited Sangdeh meteorological station and

Valikbon hydrometric station on 24 May 2010 and were able to inspect the equipment and verify

recording procedures with one of the station operators. As a result we are confident that the

Sangdeh data are probably reliable but there are thought to be some systematic errors in the data

from Valikbon resulting in proportionally larger errors as the discharge reduces (J. Parvardeh,

Tamab, pers. comm.). For the purposes of this study, however, any errors and inaccuracies in the input data will not materially affect the generalised results and interpretations of the modelling analyses.

6

Table 3. Summary of the runoff regime at Valikbon station in the Casilian Catchment. All values

are discharges in cubic metres per second. Source: Tamab.

SEASON DISCHARGE 1980 1981 1982 1983 1984 1985 1986 MEAN

Autumn highest peak 3.23 4.17 2.04 3.34 2.85 1.70 0.56 –

mean 0.41 0.40 0.44 0.34 0.29 0.22 0.11 0.32

Winter highest peak 2.22 3.46 1.78 4.15 6.25 2.20

no data –

mean 0.68 0.66 0.45 0.85 0.77 0.40 0.64

Spring highest peak 3.00 4.76 1.96 2.75 3.62 2.71 4.29 –

mean 0.47 0.63 0.59 0.73 0.71 0.43 0.74 0.64

Summer highest peak 1.92 5.75 3.63 13.87 1.88 0.73 0.66 –

mean 0.17 0.55 0.24 0.79 0.21 0.09 0.04 0.30

3.2 Climate data for future scenarios

Different GCMs have been developed in various countries of the world. For a given emissions

scenario, each GCM can be used to predict patterns of temperature and rainfall change for a region

of the Earth’s surface (Kay et al. 2009). However, GCM outputs cannot be used directly for a

specific site of interest due to their coarse scale: even in a high resolution GCM, one grid box

represents an area of greater than 50 000 km2 (Semenov et al. 1998). One method by which GCM

data can be used for small areas is to downscale the climate predictions using a “weather

generator”. This is a model that can produce synthetic weather data for continuous periods, e.g.

daily rainfall for several years, according to the statistical properties of present rainfall at a location

of interest that are modified by a proportion indicated by the GCM output.

We used the Long Ashton Research Station Weather Generator (LARS-WG) for this study

because it is readily available for use and its performance is regarded as being superior to that of

other WGs such as WGEN and Artificial Neural Network models (Semenov et al. 1998, Sajjad et

al. 2006). LARS-WG was used to downscale the Global Climate Model (GCM) outputs to the area

around Sangdeh Station to overcome the limitations of the coarse scale GCM output. A new

version, LARS-WG 5, generates rainfall and temperature data for 20-year periods 2011–2030,

2046–2065 and 2080–2099 for 13 climate models and three scenarios, having specified the

longitude, latitude and altitude of the station used to provide the statistics of the observed weather

data. The performance of this WG was evaluated by entering the daily rainfall and daily maximum

and minimum temperatures recorded at Sangdeh station from 1 January 1977 to 31 December 1996

inclusive. The software calculated the monthly mean and monthly standard deviation of these

parameters and then used these statistics to generate 300 years-worth of daily values (e.g. Semenov

et al. 1998, Sajjad et al. 2006). It then compared these statistical properties for the original and the

generated daily rainfall data using the t-test and the F-test. For both tests p < 0.05 indicating that the

means of the observed and synthetic monthly rainfall totals, and the standard deviations of the two

sets of monthly rainfall for each month, were not significantly different. Mean monthly minimum

and mean monthly maximum temperatures of the observed and generated data for all months were

identical, with only slight differences between the respective standard deviations.

To investigate the climate change effects on runoff from the Casilian Catchment, climate

data representing the effects of different climate change scenarios were required as inputs for the

rainfall-runoff model. Five different sources of uncertainty exist in climate change impact studies:

(i) future greenhouse gas emissions; (ii) GCM structure; (iii) downscaling from GCMs; (iv)

hydrological model structure; (v) hydrological model parameters (Kay et al. 2009, p.1). The

uncertainty related to the choice of GCM is greater than that from the other sources of uncertainty

7

(Bates et al. 2008, Kay et al. 2009, Boyer et al. 2010, Bae et al. 2011). GCMs contain significant

uncertainties and IPCC (2007) recommended that the results of different climate models and

scenarios should be considered in climate change studies (Abbaspour et al. 2009). Thirteen climate

models (BCM2, CNCM3, CSMK3, FGOALS, GFCM21, GIAOM, HADCM3, HADGEM, INCM3,

IPCM4, NCCCSM, NCPCM and CGMRS) and three emissions scenarios (A1B, A2 and B1) that

are available in LARS-WG provide different projections and produce different magnitudes and

patterns of rainfall and temperature changes.

Using the same observed data as for the evaluation, above, plus the coordinates of Sangdeh

station, LARS-WG produced daily rainfall and temperature data from all 13 models for three 7-year

periods (2011–2017, 2046–2052 and 2080–2086). All simulations and analyses in this study use 7-

year periods because the baseline period 1980–86 (inclusive) is the longest complete record of

discharge in the field data. The range of uncertainty in the GCM predictions (King et al. 2009,

Semenov and Stratonovitch 2010, Semenov and Shewry 2011) is indicated in Fig. 2 for annual

rainfall totals, the boxplots representing the 25th, 50th (median) and 75th percentiles of the outputs

from the thirteen GCMs with the full (1977–96 inclusive) and baseline sets of observed data

superimposed. Predicted annual rainfall totals are shown for each future simulation period under the

influence of each emissions scenario. Fig. 3 shows a similar variability of results for GCM

predictions of the number of days in each 7-year period with “heavy rainfall”, defined as daily

rainfall >95th percentile (IPCC 2007, King et al. 2009). In this study the 95th percentile from all

non-zero rainfall events for 1980–86 is 18 mm d–1

, which occurred on 50 days within this period

(i.e. seven days per year on average). The frequency of such heavy rainfall events is predicted to

increase for all future periods and scenarios in comparison with the baseline observed data.

3.3 Catchment rainfall-runoff modelling

We used the HEC-HMS rainfall-runoff model for this research. HEC-HMS is the “Hydrological

Modelling System” developed by the Hydrologic Engineering Centre of the US Corps of

Engineering and has been used successfully for many types of investigations in different parts of the

world (Cunderlik and Simonovic 2005). It is one of several rainfall-runoff modelling systems

developed in recent years and made available for application to a wide range of hydrological issues,

including investigation of the possible effects of global climate change on the hydrological

processes of a catchment.

HEC-HMS comprises two main components: a catchment model and a meteorological

model, the latter determining the net inputs to the subsurface hydrological system resulting from the

weather patterns experienced by the study location. Critical to this is the determination of the mean

monthly potential evapotranspiration (PET). The meteorological model incorporates the Penman–

Monteith method, which requires temperature, wind speed, solar radiation and relative humidity

data (Bae et al. 2011), and the Thornthwaite method which only needs temperature data (Mahdavi

2004). The latter was used in this study because only temperature data exist for the Casilian

Catchment. This was not considered a significant limitation because uncertainties in PET estimation

tend to have smaller effects on simulated runoff in climate change impact studies than the

uncertainties arising from the type of GCM or the future climate scenario (e.g. Bae et al. 2011).

HEC-HMS can also calculate runoff responses to snowfall but this requires data that do not exist for

the Casilian Catchment including snowpack characteristics and parameters that control melting (e.g.

solar radiation and dew point temperature). Consequently, it was considered unfeasible to attempt to

model snowfall for this study.

The catchment model incorporates three rainfall-runoff transformation processes. The first

of these calculates the volume of runoff from a catchment as being the proportion of rainfall

remaining in the system after “losses” have been subtracted. This is done using a 5-layer Soil

Moisture Accounting (SMA) procedure which estimates the losses to interception, infiltration,

percolation and deep percolation and subtracts them from the precipitation. These losses contribute

to canopy-interception storage, surface-depression storage, soil-profile storage and one or two

8

Figure 2. Variations in future annual rainfall projections for the Casilian Catchment produced by

LARS-WG shown as boxplots (maximum, minimum, median, upper and lower quartiles of the

mean annual values produced by 6-13 GCMs). The horizontal line represents the mean annual

rainfall for both reference periods (2.0 mm difference).

Figure 3. Variations in future projections of the number of days with “heavy rainfall”, i.e. >95th

percentile of the observed base period, within each 7-year period for the Casilian Catchment

produced by LARS-WG shown as boxplots (details as in Fig. 2). The horizontal line represents the

number of such wet days recorded during the baseline period 1980-86 inclusive.

9

layers of groundwater storage (Cunderlik and Simonovic 2005). The output from the SMA

procedure contributes to the other two transformation processes, direct runoff and groundwater

flow. The direct runoff (i.e. channel discharge) is modelled using the SCS (Soil Conservation

Service) dimensionless unit hydrograph and the groundwater flow is transformed into base flow by

a linear reservoir base flow (LRBF) method that was designed to integrate with the SMA procedure

(USACE-HEC 2000).

To model the hydrological processes in a catchment, HEC-HMS requires values for 23

parameters. Most parameters of the catchment model need only an initial estimate, their final values

being obtained through automated optimisation using the “Nelder Mead” search method. The

estimated parameter values are adjusted until the value of the selected objective function––in this

case the goodness-of-fit between the observed and computed total runoff volume from the

catchment––is minimized (Scharffenberg and Fleming 2010). For this study, initial values for all 23

parameters (listed in Table 1) were obtained from geology, soil and land use maps and from

published sources (Rawls et al. 1982, USDA 1986, Bennett 1998, McCuen 1998, Singhal and

Gupta 1999, Chainuvati and Athipanan 2001, Breuer et al. 2003, Fleming and Neary 2004, Garcı´a

et al. 2008, McEnroe 2010). A standard calibration–validation procedure was then followed (as

initially set out in Section 2.1 of Trucano et al. 2006), with parameter optimisation taking place

within the calibration stage.

4 RESULTS

4.1 Model calibration and validation

Calibration of the catchment model was performed for two different rainfall inputs, using: (i) the

mean rainfall of three stations (Sangdeh, Kale and Darzikola) calculated by the Thiessen method,

and (ii) the rainfall recorded at Sangdeh rainfall station only, to determine which option appeared to

correspond more closely with the observed runoff patterns. Observed rainfall and temperature data

were first entered into the meteorological model in order to determine the mean monthly PET, then

the catchment model was run and its calculated discharge was compared with the observed

discharge. The percentage error in the calculated annual runoff volume was designated as the

objective function for this and all subsequent model runs because this study is concerned with

future changes to runoff regimes:

% error (Rv) = (SRv – ORv) / ORv (1)

where Rv is the runoff volume, “S” indicates “simulated” and “O” is “observed”. The calibration

was performed as a “continuous” simulation using rainfall, temperature and discharge data for the

period 1 January 1980 to 31 December 1986 inclusive and the smallest error, 4.8%, was obtained

from parameter values optimised for rainfall at Sangdeh station only (Table 1). These optimised

parameter values and Sangdeh rainfall data were therefore used to validate the catchment model

using observed data from 1 January 1989 to 31 December 1992 inclusive; the resulting error of

4.9% was considered acceptable. These periods of observed data were used because they comprise

the most complete parts of the data record.

The performance of HEC-HMS was evaluated further by performing the same calibration–

validation simulations for each season separately, following Iran’s water year, but using the

optimised parameter values from the main calibration exercise. The runoff volumes for each season

were extracted from the full calibration and validation simulations and compared with the

corresponding runoff volumes recorded at Valikbon. Table 2 shows that for half of the year the

optimised parameter values do not produce very accurate results, which suggests that there are some

significant seasonal controls operating in the catchment. A strong possibility is that winter

precipitation is actually snow that accumulates rather than contributing to runoff, and that some

10

Table 1. Initial estimates of parameter values in HEC-HMS for calibration of the “continuous” model after

optimisation using the “Nelder Mead” search method and rainfall data from Sangdeh station.

PARAMETER UNITS VALUE FOR “CONTINUOUS” MODEL

Canopy storage mm 1.2

Surface storage mm 1.0

Maximum soil infiltration rate mm h–1

50.1

Impervious area % 3.6

Soil storage mm 125.0

Lag time min 1007.0

Tension storage mm 15.0

Soil percolation rate mm h–1

6.1

Initial canopy storage % 20.0

Initial surface storage % 20.0

Initial soil storage % 30.0

Initial groundwater 1 (GW1) storage % 40.0

GW1 storage mm 7.0

GW1 percolation mm h–1

0.7

GW1 coefficient h 50.2

GW1 initial discharge (linear reservoir) m3 s

–1 0.5

GW1 coefficient (linear reservoir) h 50.2

Initial groundwater 2 (GW2) storage % 40.0

GW2 storage mm 7.0

GW2 percolation mm h–1

0.7

GW2 coefficient h 50.2

GW2 initial discharge (linear reservoir) m3 s

–1 0.0

GW2 coefficient (linear reservoir) h 50.2

Table 2. Percent error in simulated runoff volumes for the full year and seasonal calibration and validation

periods.

CALIBRATION VALIDATION

Full year, 7 years

1 Jan. 1980 to 31 Dec. 1986 4.8%

Full year, 4 years

1 Jan. 1989 to 31 Dec. 1992 4.9%

CALIBRATION PERIOD (CALIBRATED) VALIDATION PERIOD (CALIBRATED)

Autumn, 7 years

23 Sep.–21 Dec. only 7.4%

Autumn, 4 years

23 Sep.–21 Dec. only 4.9%

Winter, 7 years

22 Dec.–20 Mar. only 32.7%

Winter, 4 years

22 Dec.–20 Mar. only 14.3%

Spring, 7 years

21 Mar.–21 Jun. only 21.8%

Spring, 4 years

21 Mar.–21 Jun. only 19.9%

Summer, 7 years

22 Jun.–22 Sep. only 5.2%

Summer, 4 years

22 Jun.–22 Sep. only 3.7%

runoff in spring is from snowmelt that does not correspond with any rainfall. The fact that the

calibrated model performs better against the validation data than against the calibration data for all

seasons separately may reflect inherent differences between the two datasets, such as larger

seasonal contrasts during the calibration period that would be smoothed somewhat by the model

calibration. Furthermore, we consider it very likely that some catchment parameter values vary

seasonally, particularly (mean) canopy storage relating to the deciduous tree cover but also soil

infiltration rate and soil storage, which will both be zero if the ground is frozen.

11

The absence of data relating to snowfall or the possible duration and spatial extent of

freezing conditions in the upper catchment precluded rigorous seasonal optimisation of the model

for this study. However, we did set out to explore the possible seasonal changes to runoff that may

arise in the future. Consequently we proceeded by using the calibrated parameters obtained from

annual optimisation, recognising that the seasonal results that we present must be interpreted with

caution but that they may nevertheless indicate temporal patterns that should be considered in any

future water management planning for the region.

4.2 Runoff simulations for future rainfall scenarios

Different climate models represented within LARS-WG each provide projections for every climate

scenario in terms of patterns of rainfall and temperature. The number of GCM outputs for each

scenario and future 7-year period is shown in Table 4. The variations in these outputs are shown in

Figs. 2 and 4. Fig. 4 shows the mean annual temperature already higher than the reference periods

(1980–86 or 1977–96) and increasing significantly from each 7-year period to the next in all

scenarios, though less dramatically for scenario B1. Fig. 2 indicates decreasing annual rainfall

throughout the century but with higher inter-annual variability towards the end of the century.

Again, this general trend is not as strong for scenario B1. However, the annual rainfall totals are

suggested to be initially higher than recorded during either reference period but falling to lower

levels later.

Table 4. Number of GCMs producing climate outputs for the dates and scenarios considered in this study.

A2 B1 A1B

2011-17 8 9 13

2046-52 8 9 13

2080-86 6 9 11

Table 5. Mean projected changes in annual rainfall (% of reference period values) and annual

temperature (°C) from LARS-WG and corresponding annual runoff (% of reference period values)

simulated by HEC-HMS.

Scenario: A2 B1 A1B

Dates: 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Reference

period

Percent change in annual rainfall

1980-86 5.2 0.4 -13.7 3.7 3.7 1.1 2.6 0.1 -6.9

1977-96 5.4 0.6 -13.5 3.9 3.9 1.3 2.9 0.3 -6.6

Change in annual temperature (°C)

1980-86 1.3 2.4 4.5 1.0 1.7 2.1 1.0 2.4 3.6

1977-96 1.2 2.3 4.4 0.9 1.6 2.0 0.9 2.3 3.5

Percent change in annual runoff

1980-86 4.1 -1.5 -20.6 3.4 2.7 -1.6 1.8 -2.8 -12.7

12

Figure 4. Variations in future annual temperature projections for the Casilian Catchment produced

by LARS-WG shown as boxplots (details as in Fig. 2). The upper horizontal line shows the mean

annual temperature for the 1977-96 reference period and the lower line represents the same for

1980-86 (0.1°C difference).

Figure 5. Variations in future annual runoff for the Casilian Catchment simulated by HEC-HMS

shown as boxplots (details as in Fig. 2). The horizontal line represents the mean annual runoff for

the baseline period 1980-86.

13

The median changes in rainfall and temperature obtained from the GCMs and the

corresponding changes in annual runoff calculated by HEC-HMS are indicated in Table 5 as

percentages of the mean values for the two reference periods. These annual runoff variations are

illustrated in terms of depth-equivalent runoff in Fig. 5. The patterns of projected runoff variations

are very similar to the patterns of projected annual rainfall totals. Table 6 shows the seasonal

changes as percentages of the mean values for the two reference periods. These results for winter

and spring must be considered uncertain for reasons identified earlier, although the increasing

temperatures may result in less of the winter (and spring) precipitation falling as snow with less of a

delay between precipitation and (snowmelt) runoff. However, the results in Table 6 were obtained

with no representation of snow and subsequent melting yet still show significantly higher mean

runoff in spring, an effect that may be further exaggerated by melting of winter snow and has been

highlighted by Rasoli et al. (2012) as an effect of the changing Siberian high pressure. The

predicted reduction in mean runoff in summer compared with the reference period is similarly

marked but also somewhat more reliable, notwithstanding the caveats discussed previously.

Table 5. Mean projected changes in annual rainfall (% of reference period values) and annual

temperature (°C) from LARS-WG and corresponding annual runoff (% of reference period values)

simulated by HEC-HMS.

Scenario: A2 B1 A1B

Dates: 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Reference

period

Percent change in annual rainfall

1980-86 5.2 0.4 -13.7 3.7 3.7 1.1 2.6 0.1 -6.9

1977-96 5.4 0.6 -13.5 3.9 3.9 1.3 2.9 0.3 -6.6

Change in annual temperature (°C)

1980-86 1.3 2.4 4.5 1.0 1.7 2.1 1.0 2.4 3.6

1977-96 1.2 2.3 4.4 0.9 1.6 2.0 0.9 2.3 3.5

Percent change in annual runoff

1980-86 4.1 -1.5 -20.6 3.4 2.7 -1.6 1.8 -2.8 -12.7

5 DISCUSSION

Predicted changes to annual runoff from the Casilian Catchment correspond predictably with

emissions scenarios. Scenario A2 represents slow global economic convergence with corresponding

increasing demand for energy and materials and slow development of more efficient and non-fossil

energy sources (IPCC 2000), i.e. the worst of the three represented in this study. Using all available

GCMs for the selected future periods of time and emissions scenarios incorporated into LARS-WG,

we have found that this scenario gives rise to more warming (Fig. 4) and the greatest reductions in

mean annual rainfall (Fig. 2) and, thus, runoff (Fig. 5) towards the end of the present century.

Likewise scenario B1 which represents the most optimistic case––rapid catch-up of less developed

regions but with reduced energy demand and more rapid expansion of renewable supplies (IPCC

2000)––shows the smallest changes with perhaps a very small reduction in mean annual runoff by

the 2080s. Predictions arising from scenario A1B, representing a more “balanced” pattern of global

changes between A2 and B1, fall between these extreme cases. These are, of course, very broad

generalisations that hide much variability and uncertainty. The variability among the rainfall

predictions from the different GCMs is illustrated in Fig. 2, but this integration of the different

predictions serves to reduce the degree of uncertainty inherent in each individual GCM output

(King et al. 2009, Semenov and Stratonovitch 2010, Semenov and Shewry 2011).

14

Table 6. Mean projected changes in seasonal rainfall (% of reference period values) and temperature (°C)

from LARS-WG and corresponding seasonal runoff (% of reference period values) simulated by HEC-

HMS: (a) autumn, (b) winter, (c) spring, (d) summer. In all cases, the mean value for each season from all

years in the respective data period is compared with the corresponding mean value for each season from all

years in each simulation period.

(a) A2 B1 A1B

2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Ref. period Percent change in autumn rainfall

1980-86 6.3 10.1 2.5 5.1 2.2 4.8 4.1 3.8 -3.8

1977-96 11.5 15.4 7.5 10.2 7.1 9.9 9.1 8.9 0.8

Change in autumn temperature (°C)

1980-86 1.4 2.4 4.3 0.9 1.8 2.3 1.3 2.3 3.4

1977-96 4.2 5.2 7.1 3.7 4.6 5.1 4.1 5.1 6.2

Percent change in autumn runoff

1980-86 3.8 10.0 -1.1 5.6 1.2 2.9 2.6 3.1 -8.2

(b) A2 B1 A1B

2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Ref. period Percent change in winter rainfall

1980-86 1.2 0.0 0.4 -0.9 1.2 -3.9 -0.2 -0.3 -1.9

1977-96 -2.6 -3.8 -3.4 -4.6 -2.6 -7.5 -3.9 -4.0 -5.5

Change in winter temperature (°C)

1980-86 0.9 1.9 3.4 0.6 1.5 1.6 0.7 1.9 2.7

1977-96 3.9 4.9 6.4 3.6 4.5 4.6 3.7 4.9 5.7

Percent change in winter runoff

1980-86 1.2 -2.3 -3.0 -2.3 -0.8 -4.8 -2.0 -2.1 -4.1

(c) A2 B1 A1B

2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Ref. period Percent change in spring rainfall

1980-86 18.5 9.7 -7.5 22.1 13.3 12.8 17.7 11.2 -0.1

1977-96 7.9 -0.1 -15.8 11.1 3.1 2.6 7.1 1.2 -9.1

Change in spring temperature (°C)

1980-86 1.2 2.5 4.6 1.0 1.8 2.3 1.0 2.5 3.4

1977-96 4.3 5.6 7.7 4.1 4.9 5.4 4.1 5.6 6.5

Percent change in spring runoff

1980-86 32.3 14.2 -5.3 32.4 22.6 18.8 25.9 18.6 3.8

(d) A2 B1 A1B

2011-17 2046-52 2080-86 2011-17 2046-52 2080-86 2011-17 2046-52 2080-86

Ref. period Percent change in summer rainfall

1980-86 -12.9 -20.4 -31.1 -8.6 -10.2 -14.9 -11.1 -19.1 -22.3

1977-96 -9.9 -17.7 -28.7 -5.4 -7.1 -12.0 -8.0 -16.3 -19.6

Change in summer temperature (°C)

1980-86 1.3 2.7 5.1 1.2 2.0 2.6 1.0 2.7 3.8

1977-96 4.1 5.5 7.9 4.0 4.8 5.4 3.8 5.5 6.6

Percent change in summer runoff

1980-86 -18.0 -28.5 -41.8 -12.9 -15.7 -21.0 -15.9 -23.6 -30.1

Predicted changes to seasonal runoff from the Casilian Catchment show broadly the same

patterns across the three scenarios but future change compared with the reference periods may be

insignificant in autumn and winter. However, the modelling suggests rainfall and total runoff to be

higher in spring and much lower in summer. Using LARS-WG, all scenarios suggest that present

15

spring rainfall (i.e. 2011–17) should be ~20% higher than in 1980–86 or ~8–10% higher than 1977–

96, and that present summer rainfall should be ~10% lower than in 1980–86 and 1977–96 (Table 6).

A study of rainfall data from 28 synoptic stations in Iran over the period 1967–2006 found that in

northern Iran rainfall changed by +0–10% annually, +10% in spring, +0–30% in summer, +0–10%

in autumn and -10–20% in winter (ShiftehSome'e et al. 2012). The predictions for autumn and

spring thus appear consistent with observations but those for summer and winter do not. However,

fifteen recent AOGCM runs for a future warmer climate indicate a decrease in summer rainfall in

most parts of the mid-latitudes (which include Iran), indicating a greater risk of droughts in these

regions in summer (Bates et al. 2008, p.38).

The uncertainties associated with the prediction of future climates are well documented, not

least arising from the widely varying patterns of rainfall generated by different climate models

(Bates et al. 2008). Previous studies have explored the possible effects of climate change on water

resources in Iran (Babaeian et al. 2007; Abbaspour et al. 2009; Zarghami et al. 2011) and elsewhere

(e.g. Bae et al. 2008) but using outputs from only one GCM, which greatly limits the validity and

usefulness of the findings. More generally, GCMs have projected that mean water vapour,

evaporation, mean rainfall especially mean annual rainfall, extreme rainfall events and rainfall

intensity are likely to increase in the future (Karl et al. 1996, SWCS 2003, IPCC 2007, Zhang et al.

2009). Indeed, a small increase in mean annual rainfall that has been observed worldwide during the

20th century has been the result of heavy rainfall events (Easterling et al. 2000, Nearing et al. 2005,

Rahimzadeh 2009). For example, Karl and Knight (1998) and Groisman et al. (2001) reported an

increase by 10% from 1910 to 1996 over the United States with 53% of this increase arising from

10% of precipitation events (i.e. the most intense rainfall) (Nearing et al. 2005).

Drought and flooding are the two high risk Iranian weather conditions. The predicted

decrease in summer rainfall and higher evapotranspiration (due to higher temperatures) can be

expected to cause a reduction in the soil water and groundwater recharge. Indeed, even with higher

mean rainfall, there may be less groundwater recharge if the rain falls primarily during high

intensity convective events that generate surface runoff rather than infiltrating into the soil. Thus,

any water management plans for northern Iran must take account of both types of hazard, such as an

integrated proposal that floodwater arising from higher spring rainfall should be stored using

different methods including ground water recharge systems and recharge ponds (Sharifi et al. 2012)

to increase the availability of water during drought periods.

Irrespective of the seasonal patterns, Figure 3 suggests that whichever scenario applies, the

frequency of “extreme events” in the Casilian Catchment (i.e. >18 mm d–1

) will increase compared

with the baseline period. In general, more frequent extreme events in the future can be expected to

result in more flooding events for any catchment at any time of year (Abbaspour et al. 2009),

particularly in the form of “flash floods” and often related rainfall-induced hazards such as debris

flows and other types of landslides. Detailed hazard and risk assessment exercises will be needed in

order to devise appropriate strategies for minimising and mitigating the impacts of these types of

events. Typical strategies may include hazard zonation, for example, preventing further

development in high risk locations, and updating infrastructure specifications such as larger culverts

and bridge spans to accommodate higher discharges and reduce accumulation of floodwater

upstream of such structures. As always, the implementation of such strategies is likely to depend on

perceptions of risk as well as scientific interpretations of probabilities balanced against local and

regional economic and political factors including the healthy tourism industry in northern Iran.

6 CONCLUSIONS

This study suggests that future patterns of runoff from forested catchments bordering the south

coast of the Caspian Sea in northern Iran will directly correspond with predicted changes in rainfall.

Overall, patterns of rainfall and runoff in autumn and winter may be little different from the present,

but that significantly higher rainfall and runoff may occur during spring with significantly lower

16

rainfall and runoff in summer. The corresponding hazard scenarios are thus for higher probabilities

of regional-scale flooding in spring and droughts in summer. This presents a fortuitously ideal basis

for integrated water management strategies that capture and store excess spring runoff in order to

mitigate the effects of possible summer droughts. However, a higher proportion of the rainfall is

expected to occur as occasional high intensity events that can be expected to generate flash floods–

and possibly landslides – at any time of year. Management and mitigation of these hazards will be

required, particularly if the tourism industry of northern Iran is not to be adversely affected in the

future.

Acknowledgements We are particularly grateful to the late J. Parvardeh, Head of the Surface Water

Section in Tamab (Water Resources Researches Organization of Iran), who provided relevant

catchment data to BZ and then FH over many years, and to staff of the Farim Wood Company who

provided access for FH, BZ and APD to Sangdeh and Valikbon monitoring stations. We also

acknowledge individuals and organisations who assisted FH with his research in Tehran: National

Cartographic Centre of Iran; Geological Survey of Iran; Dr Saeed Morid (Forests, Range and

Watershed Management Organisation, Tehran, and Tarbiate Modares University); Dr Rahim Salavi-

Tabar (Mahab Ghodss Consulting Engineering Company); Moshanir Power Engineering

Consultants; Dr Iman Babaeian (Atmospheric Science and Meteorological Research Centre,

Tehran); Dr Niaz Ali Ebrahimipak (Soil and Water Research Institute, Tehran); and, in the UK, Dr

Mikhail Semenov (Rothamsted Research, Harpenden).

Funding This research was funded by a scholarship from Azad University in Tehran to the first

author to undertake a PhD at Kingston University, UK.

Disclosure Statement There are no financial interests or benefits arising from this work.

References

Abbaspour, K.C., Faramarzi, M., Ghasemi, S.S. and Yang, H., 2009. Assessing the impact of

climate change on water resources in Iran. Water Resources, 45, W10434, doi:

10.1029/2008WR007615.

Babaeian, I., Najafinik, Z., Zabolabasi, F., Habibinokhandan, M., Adab, H. and Malbosi, S., 2007.

Climate change investigation of Iran for 2010-2039 using downscaled output of ECHO-G

climate model. Geography and Development Iranian Journal, 16, 135–152 (in Persian).

Bae, D.H., Jung, W. and Chang, H., 2008. Potential changes in Korean water resources estimated

by high-resolution climate resolution. Climate Research, 35, 213–226.

Bae, D.H., Jung, W. and Lettenmaier, D.P., 2011. Hydrologic uncertainties in climate change from

IPCC AR4 GCM simulations of the Chungju Catchment, Korea. Journal of Hydrology, 401, 90–

105.

Barnett, D.N., Brown, S.J., Murphy, J.M., Sexton, D.M.H. and Webb, M.J., 2006. Quantifying

uncertainty in changes in extreme event frequency in response to doubled CO2 using a large

ensemble of GCM simulations. Climate Dynamics, 26, 489–511.

Banasik, K. and Hejduk, L., 2012. Long-term changes in runoff from a small agricultural

catchment. Soil and Water Research, 7, 64–72.

Bates, B.C., Kundzewicz, Z.W., Wu, S. and Palutikof, J.P., eds., 2008. Climate change and water.

Technical paper of the Intergovernmental Panel on Climate Change. Geneva: IPCC Secretariat.

Bennet, T., 1998. Development and application of a continuous Soil Moisture Accounting

algorithm for the Hydrologic Engineering Center Hydrologic Modeling System (HEC-HMS).

Unpublished MS thesis. University Of California (Davis).

17

Boyer, C., Chaumont, D., Chartier, I. and Roy, A.G., 2010. Impact of climate change on the

hydrology of St. Lawrence tributaries. Journal of Hydrology, 384, 65–83.

Breuer, L., Eckhardt, K. and Frede, H.G., 2003. Plant parameter values for models in temperate

climates. Ecological Modelling, 169, 237–293.

Chainuvati, C. and Athipanan, W., 2001. Crop diversification in Thailand. In: M.K. Papademetriou

and F.J. Dent, eds. Crop diversification in the Asia–Pacific region [online]. Bangkok: Food and

Agriculture Organization. RAP publication 2001/03. Available from:

www.fao.org/docrep/003/x6906e/x6906e00.htm#contents(Accessed 15 July 2015).

Chang, J., Wang, Y., Istanbulluoglu, E., Bai, T., Huang, Q., Yang, D. and Huang, S., 2015. Impact

of climate change and human activities on runoff in the Weihe River Basin, China. Quaternary

International, 380–381,169–179.

Cunderlik, J.M. and Simonovic, S.P., 2005. Hydrological extremes in a southwestern Ontario river

basin under future climate conditions/Extrêmes hydrologiques dans un basin versant du sud-ouest

de l'Ontario sous conditions climatiques futures. Hydrological Sciences Journal, 50 (4), 631–654.

Douville, H., Chauvin, F., Planton, S., Royer, J.F., Salas-Melia, D. and Tyteca, S., 2002. Sensitivity

of the hydrological cycle to increasing amounts of greenhouse gases and aerosols. Climate

Dynamics, 20, 45–68.

Easterling, D.R., Karl, T.R., Gallo, K.P., Robinson, D.A., Trenberth, K.E. and Dai, A., 2000.

Observed climate variability and change of relevance to the biosphere. Journal of Geophysical

Research, 105 (D15), 101–114.

Fleming, M. and Neary, V.S., 2004. Continuous hydrologic modeling study with Hydrologic

Modelling System (HEC-HMS). Journal of Hydrologic Engineering, 9(3), 175–183.

Garcı´a, A., Sainz, A., Revilla, J.A., A´lvarez, C., Juanes, J.A. and Puente, A., 2008. Surface water

resources assessment in scarcely gauged catchments in the north of Spain. Journal of Hydrology,

356, 312–326.

Groisman, P.Y., Knight, R.W. and Karl, T.R., 2001. Heavy precipitation and high stream flow in

the contiguous United States: Trends in the twentieth century. Bulletin of the American

Meteorological Society, 82 (2), 219–246.

IPCC, 2000. Emissions Scenarios [N. Nakicenovic and R. Swart R, eds.] Cambridge University

Press: Intergovernmental Panel on Climate Change.

IPCC, 2007. Climate Change 2007, The Physical Science Basis. Contribution of Working Group I

to the Fourth Assessment Report. [S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B.

Averyt, M. Tignor and H.L. Miller, eds.] Cambridge University Press: Intergovernmental Panel

on Climate Change.

Jiang, C., Zhang, L., Li, D. and Li, F., 2015. Water discharge and sediment load changes in China:

Change patterns, causes, and implications. Water, 7, 5849–5875, doi:10.3390/w7105849.

Karl, T.R. and Knight, R.W., 1998. Secular trends of precipitation amount, frequency, and intensity

in the USA. Bulletin of the American Meteorological Society, 79, 231–241.

Karl, T.R., Knight, R.W., Easterling, D.R. and Quayle, R.G., 1996. Indices of climate change for

the United States. Bulletin of the American Meteorological Society, 77, 279–292.

Katiraie, P.S., Hajjam, S. and Irannejad, P., 2007. Contribution from the variations of precipitation

frequency and daily intensity to the precipitation trend in Iran over the period 1960–2001.

Journal of the Earth and Space Physics, 33(1), 67–83 (in Persian).

Kay, A.L., Davies, H.N., Bell, V.A. and Jones, R.G., 2009. Comparison of uncertainty sources for

climate change impacts: flood frequency in England. Climatic Change, 92, 41–63.

Kharin, V.V. and Zwiers, F.W., 2005. Estimating extremes in transient climate change simulations.

Journal of Climate, 18, 1156–1173.

King, L., Solaiman, T. and Simonovic, S.P., 2009. Assessment of climatic vulnerability in the Upper

Thames river catchment [online]. London (Ontario): Department of Civil and Environmental

Engineering, University of Western Ontario, Water Resources Research Report No. 64.

Available from:www.eng.uwo.ca/research/iclr/fids/publications/products/66.pdf (Accessed 16

July 2015).

18

Kundzewicz, Z.W., Kanae, S., Seneviratne, S.I., Handmer, J., Nicholls, N., Peduzzi, P., Mechler,

R., Bouwer, L.M., Arnell, N., Mach, K., Muir-Wood, R., Brakenridge, G.R., Kron, W., Benito,

G., Honda, Y., Takahashi, K. and Sherstyukov, B., 2013. Flood risk and climate change: global

and regional perspectives. Hydrological Sciences Journal, 59, 1–28.

Langhammer, J., Su, Y. and Bernsteinová, J., 2015. Runoff response to climate warming and forest

disturbance in a mid-mountain basin. Water, 7, 3320–3342, doi:10.3390/w7073320.

Li, F., Zhang, G. and Xu, Y.J., 2014. Separating the impacts of climate variation and human

activities on runoff in the Songhua River Basin, Northeast China. Water, 6, 3320–3338,

doi:10.3390/w6113320.

Mahdavi, M., 2004. Applied hydrology. University of Tehran Press (in Persian).

McCuen, R.H., 1998. Hydrologic analysis and design. Upper Saddle River (NJ): Prentice Hall.

McEnroe, B.M., 2010. Guidelines for continuous simulation of stream flow in Johnson County,

Kansas, with HEC-HMS [online]. Lawrence: Department of Civil, Environmental and

Architectural Engineering, University of Kansas, report to Johnson County public works and

infrastructure storm water management program. Available from:

http://www.jocogov.org/sites/default/files/documents/PWK/continuous-simulation-with-HEC-

HMS.pdf (Accessed 16 July 2015).

Nearing, M., Jetten, V., Baffaut, C., Cerdan, O., Couturier, A., Hernandez, M., Bissonnais, Y.,

Nichols, M., Nunes, J., Renschler, C., Souchere, V. and van Oost, K., 2005. Modelling response

of soil erosion and runoff to changes in precipitation and cover. Catena, 61, 131–154.

Parvardeh, J., 1985. Data and studies of surface waters. Unpublished report of the Surface Water

Section of the Water Resources Researches Organization of Iran (TAMAB). (In Persian).

Rahimzadeh, F., Asgari, A. and Fattahi, E., 2009. Variability of extreme temperature and

precipitation in Iran during recent decades. International Journal of Climatology, 29, 329–343.

Rasoli, A., Babaeian, I., Ghaemi, H. and Zevarreza, P., 2012. Time series analysis of pressure of

centre of synoptic patterns influencing the seasonal rainfall of Iran. Geography and

Development, 27, 77–88 (in Persian).

Rawls, W.J., Brakensiek, D.L. and Saxton, K.E., 1982. Estimation of soil water properties.

Transactions of the American Society of Agricultural Engineers, 108, 1316–1320.

Sajjad Khan, M., Coulibaly, P. and Dibike, Y., 2006. Uncertainty analysis of statistical downscaling

methods. Journal of Hydrology, 319, 357–382.

Scharffenberg, W.A. and Fleming, M.J., 2010. Hydrologic Modelling System HEC-HMS User’s

Manual. Davis (California): US Army Corps of Engineers.

Semenov, M.A. and Stratonovitch, P., 2010. Use of multi-model ensembles from global climate

models for assessment of climate change impacts. Climate Research, 41, 1–14.

Semenov, M. and Shewry, P.R., 2011. Modelling predicts that heat stress, not drought, will increase

vulnerability of wheat in Europe. Scientific Reports, 1, 66.

Semenov, M.A., Brooks, R.J., Barrow, E.M. and Richardson, C.W., 1998. Comparison of the

WGEN and LARS-WG stochastic weather generators in diverse climates. Climate Research, 10,

95–107.

Sharifi, F., Samadi, S.Z. and Wilson, C.A.M.E., 2012. Causes and consequences of recent floods in

the Golestan catchments and Caspian Sea regions of Iran. Natural Hazards, 61, 533–550.

ShiftehSome'e, B., Ezani, A. and Tabari, H., 2012. Spatiotemporal trends and change point of

precipitation in Iran. Atmospheric Research, 113, 1–12.

Singhal, B.B.S. and Gupta, R.P., 1999. Applied hydrogeology of fractured rocks. Dordrecht:

Kluwer Academic.

Stahl, K., Hisdal, H., Hannaford, J., Tallaksen, L.M., van Lanen, H.A.J., Sauquet, E., Demuth, S.,

Fendekova, M. and Jódar, J., 2010. Streamflow trends in Europe: evidence from a dataset of

near-natural catchments. Hydrology and Earth System Sciences, 14, 2367–2382,

doi:10.5194/hess-14-2367-2010.

19

SWCS, 2003. Conservation implications of climate change: soil erosion and runoff from cropland

[online]. Ankeny (Iowa): Soil and Water Conservation Society. Available from:

http://www.swcs.org/documents/filelibrary/advocacy_publications_before_2005/Climate_chang

efinal_112904154622.pdf (Accessed 16 July 2015).

Tabari, H., Nikbakht, J. and ShiftehSome’e, B., 2011. Investigation of groundwater level

fluctuations in the north of Iran. Environmental Earth Sciences, doi:10.1007/s12665-011-1229-z.

Trucanoa, T.G., Swilera, L.P., Igusab, T., Oberkampfc, W.L. and Pilchc, M., 2006. Calibration,

validation, and sensitivity analysis: What’s what. Reliability, Engineering and System Safety, 91,

1331–1357.

USACE-HEC, 2000. Hydrologic Modelling System, HEC-HMS: Technical Reference Manual.

Davis (California): US Army Corps of Engineers, Hydrologic Engineering Centre CPD-74B.

USDA, 1986. Urban hydrology for small catchments [online].Washington D.C.: United States

Department of Agriculture, Technical Release 55. Available from:

http://cset.sp.utoledo.edu/~nkissoff/pdf/CIVE-3520/Modified-tr55.pdf (Accessed 16 July 2015).

Venturini, V., Islam, S. and Rodriguez, L., 2008. Estimation of evaporative fraction and

evapotranspiration from MODIS products using a complementary based model. Remote Sensing

of the Environment, 112, 132–141.

Wilby, R.L. and Wigley, T.M.L., 2002. Future changes in the distribution of daily precipitation

totals across North America. Geophysical Research Letters, 29(7), 1135.

Zampieri, M., Scoccimarro, E., Gualdi, S. and Navarra, A., 2015. Observed shift towards earlier

spring discharge in the main Alpine rivers. Science of the Total Environment, 503–504, 222–232.

Zarghami, M., Abdi, A., Babaeian, I., Hassanzadeh, Y. and Kanani, R., 2011. Impacts of climate

change on runoffs in East Azerbaijan, Iran. Global and Planetary Change, 78(3–4), 137–146.

Zhan, C.S., Jiang, S.S., Sun, F.B., Jia, Y.W., Niu, C.W. and Yue, W.F., 2014. Quantitative

contribution of climate change and human activities to runoff changes in the Wei River basin,

China. Hydrology and Earth System Sciences, 18, 3069–3077, doi:10.5194/hess-18-3069-2014.

Zhang, X.C., Liu, W.Z., Li, Z. and Zheng, F.L., 2009. Simulating site-specific impacts of climate

change on soil erosion and surface hydrology in southern Loess Plateau of China. Catena, 79,

237–242.